Hi all. When ranked mm first came out I calibrated at 3.7k. I'm now 5.5k solo, though I've been hovering in and out of the leaderboards for the last week. I thought I might share some of my thoughts on gaining mmr that might be helpful/new to many of you. Everything here refers to solo queue, though most of it's also applicable to playing with friends.
MMR is a reflection of someone's overall skill level
However, given how complex of a game dota is, it's a relatively coarse measurement. Some people might say MMR doesn't mean anything, and they would be wrong. It is, by definition (ie, you gain mmr when you win and you lose mmr when you lose), a measurement of how capable you are at winning games.As a reflection of your skill level, you can't increase your mmr in the long term (ie, ignoring short term fluctuations due to randomness) without actually becoming better at winning games
This sentence is a little vague, and some ways of making the sentence precise are false, but essentially the philosophy behind this, is that in order to increase your mmr you must either A) Become a better player, or B) Make your team play better.- I will first consider option (A) above - ie, getting better at dota. Now, there are tons of guides applicable to all sorts of skill ranges that discuss this in varying levels of detail, so what I'm going to focus on aren't the details. I'm going to consider the big picture, which is.
So how do you get better at ANYTHING?
Suppose you want to learn to play Chopin's Ballade in G minor. This is a very difficult piece of music for the piano, and is about 8-10 pages long (iirc). If you want to get better at playing this, you could begin by sightreading it and trying to play the entire thing extremely slowly, which could take you as much as 20-30 minutes depending on how much you stumble. And then maybe you could do it again, and again, and again. Eventually you'll get better at the piece, but this would be an extremely inefficient way of learning the piece. Why? Because your brain doesn't have enough working memory to store the entire playthrough of the piece, along with all the little details and bits that you stumbled on, so if you stumbled on a part near the beginning, by the time you've reached the end 20 minutes later, you've probably forgotten about that stumble near the beginning, and it becomes difficult to use the memory of that stumbling to get better at playing that part in future playthroughs. Instead, what pretty much every piano student does, is to divide the piece into parts, and practice each part thoroughly before moving on to the next part. The parts may vary in length between only a few measures to an entire page, with the most difficult parts being the shortest in length (so you can play them as many times as possible in a reasonable practice session). The same philosophy holds in dota. It's really difficult to just become a better dota player without focusing on the areas you need to improve on. However, instead of dividing up dota into aspects like last hitting, map awareness, etc. I think it's just as reasonable to divide it into heroes. If you want to get better at dota, what I would advocate, is picking a small list of heroes (maybe 2-4 heroes) that you just focus on getting better at, and play them all 10-20 times each. This takes us to the next step.SELF-ASSESSMENT... is hard, especially with such a complex game as dota
In fact, assessing how well you're doing, what you need to improve on, is difficult no matter what you're trying to learn. This is why people hire teachers/go to class. Now, while I think hiring a high-mmr player to coach you will definitely help you get better, I understand many dota players are poor and don't want to use what little cash they have this way (buying hats is obviously a much better use of your money o.0). Honestly, I can't tell you how to best analyze your own play. There are so many things to consider for that to write them here would make this post literally unreadable. However, I can say that honestly the key to getting better, is to NEVER go into a game hoping to win. Winning is great, but ultimately it is a side effect to playing well, and your goal going into every game is to play well. You should constantly be thinking about what your best next move is. Sometimes you will need to coordinate with your team, but every time you try to coordinate you should know that there's a chance they don't cooperate, in which case you should have a backup plan for what you should do if they don't cooperate. If you think you should gank the enemy carry but you need the help of your mid, you can ask him to help you, but if he says no or doesn't respond, going in solo would be stupid, so have a backup plan! (which could even be to just farm).Why shouldn't you go into a game hoping to win?
Because it's too easy to spend anywhere from 10 seconds to a few minutes dwelling on your allies mistakes (every time they happen) and pondering your own misfortune at being teamed up with these idiots instead of thinking about what you should be doing right now. For me, often when I hit a record mmr I become terrified of losing it, and as a result the next time I play a game and someone on my team (or even me) makes a mistake I might spend 30s arguing with them if it was their mistake or defending myself if it was my mistake, during which I just go into autopilot and stop thinking about how to best spend my time. I can't count the number of times this has caused me to lose focus, not watch the map as closely, and get ganked, exacerbating the problem. If you don't focus on winning, but instead on playing well recognizing the fact that there's roughly a 50% chance of you losing this game, then it becomes easier to just note the mistakes that were made, move on, and maintain your focus. Plus, this way, it's easier to look back on every game and feel good about the lessons you've learned that game. Whether you won or lost is irrelevant. If you're learning from mistakes then you're becoming a better player, and your mmr will naturally increase in the long term as a result. To this end, if you don't have a tough skin, don't be afraid to mute people who flame you. Personally in the loading screen of every game, I make sure I unmute everyone. Then, once the game starts, as soon as someone flames me I tell them if they keep flaming me in a nonconstructive manner I'm going to mute them, and I follow through with the promise. Yes dota is a team game and communication is important to winning, but in the end your goal should be to focus on your own play and play well, and if your teammates are making that difficult for you then muting them is the obvious solution (and may make you end up winning the game anyway thanks to your better play).The next goal to have when going into a game is to use the data you get from the game to revise the model of the game that you have in your head
I can't count the number of times I've gotten doomsayers on your team saying "omg that AM is so farmed its gg", or "gg no gank spectre". Then, when the teamfights roll in their AM or specter gets hexed and blown up or somehow get kited or don't do enough damage and you win. If you find yourself saying something like that and then still end up winning the game, you should really think about at the end of the game why you were wrong. Honestly, there could be a thousand reasons. Generally this comes in two flavors. Firstly, maybe your team actually had good ways to deal with the enemy no matter how farmed they get, in which case identifying the factors that let your team deal with the enemy carry should be very valuable in making the model of dota you have in your head more accurate. Secondly, maybe your enemy carry made a mistake. Now this is much harder to identify, and unless you're a pro a lot of things you might consider mistakes actually aren't mistakes, and vice versa. Either way, even if he actually made a mistake, you should think - why did he make this mistake? What was he thinking? Is this a mistake that is often made at your mmr? For example, it's very common in mmrs below 4k to get absolute freefarm for the first 15 minutes and to think that you can just go ham on the enemy as a 1100 hp luna with treads drums midas dominator, without realizing that 1100 hp isn't actually very much and in a teamfight without a bkb you can easily get disabled and focused down before you can even press R (in fact it's an interesting aspect of dota that given two teams with exactly the same hero compositions and exactly the same items/levels, that teamfights will often be stomps in one way or the other - ie, good initiations and vision advantages via high grounds/wards is incredibly important). Anyway, if they made a mistake that is often made, then that's something you can exploit in the future when playing against a farmed <whatever>, and something you should note not to do when playing as a farmed <whatever>. In fact, if you understand why he made the mistake, you can even try to bait out the mistake by creating similar conditions in the future. For example, if 3v3 fight breaks out somewhere on the map (especially at a tower), often the other 4 heroes might watch the fight and try to decide if they should tp to help or keep farming. In that situation, the enemy carry might take an extra second to react to you trying to gank him with the 2 heroes on your team not in that fight and as a result die due to not backing up in time. Was that a mistake? Yes! Did you get lucky that he was slow to escape? Yes! But the point was that you made your own luck by timing the gank to coincide with a fight that caused him to look away from his hero.Of course, the above point relies on having accurate data
You can't improve your theory/model of how dota works without accurate data, which is sometimes difficult to collect in game, since things often happen so fast. However, thankfully there is such a thing as the amazing downloadable replay. If you're really serious about getting better at dota, then watching your replays is an absolute necessity. Knowing what to look for in a replay is difficult and I can't really write everything down, but a few things to consider are: How did this fight start? Were we looking for a fight? Should we have been looking for a fight? If we won/lost, did I think we were going to win/lose at the beginning of the fight? How does this fit with my theory of dota? Here, the last question is especially important, as often in mmrs below 4k, the vast majority of fights are actually just a cacophony of fuckups. Everyone misses their spells or uses them on the wrong targets, people run when they should be chasing or chase when they should be running, so it's incredibly important to understand what mistakes were made (especially by you), and to understand if the fight should have been a win or a loss. Of course, even most <5k players won't be able to actually answer these questions correctly, but at least thinking about the question allows you to have a basepoint from which you can adjust your answers/theory of dota in light of more data from future games and similar fights.I'm getting a bit tired now and so I'll end with this last point, which is - how do you make your team play better?
(ie, (B) from point 2) A lot of people are emotional when they play dota, and the variation in the quality of someone's play from game to game in part can be attributed to their state of mind. People enjoy the feeling of being a part of a cooperative team who's got each others backs, so if you can somehow instill that feeling in your teammates, you're actually improving your ability to win the game without actually being any better at dota yourself (this is the second way of gaining mmr!) There are a number of ways to do this. Essentially, you just want to be serious, but friendly. You want to limit your communication to things that are useful to your teammates, such as missing calls, pinging runes, noting which enemy heros are carrying wards, where they placed them...etc. You should never flame allies. It will only make them play worse. You can point out mistakes your allies made only if you are positive it was a mistake, and if you think they are likely to make the same mistakes again in the same game, but you want to be as specific as possible, and never phrase it in an aggressive tone. Give your supports farm. A 2k networth support at 30 minutes is actually worthless. If you're a core hero feel free to tell your supports to take a lane when you're off ganking or when you're strong enough to jungle. This applies to other cores too. Sometimes after I start a tp to a lane to farm the incoming creep wave, a teammate will still tp there after me. That's obviously a waste of time for us, so I would generally say something like "can I have this please", and if he still tries to compete for the last hits, I'll just move into the jungle or do something else.Lastly, here's a trick I often use : At the beginning of the game, if I sense a bad atmosphere for whatever reason, I will find a reason to commend some player on my team (and actually commend them, and announce that you're commending them for <whatever>). The reason should only have to do with their attitude, not their skill, and ideally you should do this as early as possible. Everyone likes to be commended, so the person you commended will probably like you more, hence improving the team dynamics between you two, but more importantly, the people who weren't commended are now aware that you actually commend people, and as a result may subconsciously change their behavior in a way to also deserve a commend from you. For example, last night when I noticed that my game was on page 22 and hence I was probably playing with a bunch of 4k players, after our solo support said depressingly "great, I'm solo support again...", and proceeded to buy courier and wards, I instantly commended him and announced "commended teal for his attitude". That was honestly the best game I've ever had with a bunch of 4k strangers (he also continued to buy support items in a timely manner for the rest of the game!). Even if 3 of your teammates are already flaming each other right out of the gate, you can even commend the person who didn't say anything for "keeping a cool head" or something.
TLDR:
- Focus on your own play. Never go into a game hoping to win. Always go into a game trying to play well and learn from mistakes.
- Keep a model/theory of dota in your head. Constantly think about how your experiences in games fits with that model, and adjust it accordingly. Watch replays.
- Happy teammates = better teammates. You can win games solely off of making your teammates feel good. Be friendly, but serious. Never flame. Make constructive criticism if you think it will help. Commend people for their attitude.
by kappaoxeimon at Reddit